Intravital video microscopy was used to investigate leukocyte margination in 80 mesenteric venules (19 to 54 microns) of 50 anesthetized rabbits. After intravenous (IV) bolus injection, sulfated polysaccharides reduced in a reversible and dose-dependent way the number of leukocytes rolling slowly along the venular wall. The presence of sulfate groups is essential because other negatively charged or neutral polysaccharides had no effect. It was not caused by an increase in RBC velocity or chelation of divalent cations. Inhibition by sulfated dextrans (n = 7) was independent of molecular weight (mol wt 13,000 to 500,000) but was influenced by the average number of sulfate groups per monosaccharide. With substitution 0.13, the 90%-inhibition dose was 104 mg/kg, with 0.7 it was 56 mg/kg, and between substitution 1 and 2 it ranged from 20 to 23 mg/kg. At 100 mg/kg, plasma concentration was 0.6 to 0.7 mg/mL. Xylan sulfate (mol wt 6,000, substitution 1.8) gave 90% inhibition at 11 mg/kg, and heparin gave 90% inhibition at 97 mg/kg. Duration of inhibition (0.5 to 2 hours) depended on mol wt and appeared to be related to plasma clearance. Because protamine also inhibited rolling (12 mg/kg; less than 10 minutes), we propose that repetitive formation and breakup of ionic bonds between sulfate groups and positively charged amino acids is involved in leukocyte rolling. During inhibition of rolling, systemic lymphocyte/monocyte levels appeared to increase. Granulocyte counts did not change, indicating that rolling is not the main mechanism responsible for the marginal granulocyte pool.